I. Ions, Electric Fields and Fluid Flow: Field-induced Formation of Planar Bead Arrays
Electrokinesis refers to a class of phenomena elicited by the action of an electric field on the mobile ions surrounding charged objects in an electrolyte solution. When an object of given surface charge is immersed in a solution containing ions, a diffuse ion cloud forms to screen the object's surface charge. This arrangement of a layer of (immobile) charges associated with an immersed object and the screening cloud of (mobile) counterions in solution is referred to as a “double layer”. In this region of small but finite thickness, the fluid is not electroneutral. Consequently, electric fields acting on this region will set in motion ions in the diffuse layer, and these will in turn entrain the surrounding fluid. The resulting flow fields reflect the spatial distribution of ionic current in the fluid. Electroosmosis represents the simplest example of an electrokinetic phenomenon. It arises when an electric field is applied parallel to the surface of a sample container or electrode exhibiting fixed surface charges, as in the case of a silicon oxide electrode (in the range of neutral pH). As counterions in the electrode double layer are accelerated by the electric field, they drag along solvent molecules and set up bulk fluid flow. This effect can be very substantial in narrow capillaries and may be used to advantage to devise fluid pumping systems.
Electrophoresis is a related phenomenon which refers to the field-induced transport of charged particles immersed in an electrolyte. As with electroosmosis, an electric field accelerates mobile ions in the double layer of the particle. If, in contrast to the earlier case, the particle itself is mobile, it will compensate for this field-induced motion of ions (and the resulting ionic current) by moving in the opposite direction. Electrophoresis plays an important role in industrial coating processes and, along with electroosmosis, it is of particular interest in connection with the development of capillary electrophoresis into a mainstay of modern bioanalytical separation technology.
In confined geometries, such as that of a shallow experimental chamber in the form of a “sandwich” of two planar electrodes, the surface charge distribution and topography of the bounding electrode surfaces play a particularly important role in determining the nature and spatial structure of electroosmotic flow. Such a “sandwich” electrochemical cell may be formed by a pair of electrodes separated by a shallow gap. Typically, the bottom electrode will be formed by an oxide-capped silicon wafer, while the other electrode is formed by optically transparent, conducting indium tin oxide (ITO). The silicon (Si) wafer represents a thin slice of a single crystal of silicon which is doped to attain suitable levels of electrical conductivity and insulated from the electrolyte solution by a thin layer of silicon oxide (SiOx).
The reversible aggregation of beads into planar aggregates adjacent to an electrode surface may be induced by a (DC or AC) electric field that is applied normal to the electrode surface. While the phenomenon has been previously observed in a cell formed by a pair of conductive ITO electrodes (Richetti, Prost and Barois, J. Physique Lettr. 45, L-1137 through L-1143 (1984)), the contents of which are incorporated herein by reference, it has been only recently demonstrate that the underlying attractive interaction between beads is mediated by electrokinetic flow (Yeh, Seul and Shraiman, “Assembly of Ordered Colloidal Aggregates by Electric Field Induced Fluid Flow”, Nature 386, 57-59 (1997), the contents of which are incorporated herein by reference. This flow reflects the action of lateral non-uniformities in the spatial distribution of the current in the vicinity of the electrode. In the simplest case, such non-uniformities are introduced by the very presence of a colloidal bead near the electrode as a result of the fact that each bead interferes with the motion of ions in the electrolyte. Thus, it has been observed that an individual bead, when placed near the electrode surface, generates a toroidal flow of fluid centered on the bead. Spatial non-uniformities in the properties of the electrode can also be introduced deliberately by several methods to produce lateral fluid flow toward regions of low impedance. These methods are described in subsequent sections below.
Particles embedded in the electrokinetic flow are advected regardless of their specific chemical or biological nature, while simultaneously altering the flow field. As a result, the electric field-induced assembly of planar aggregates and arrays applies to such diverse particles as: colloidal polymer lattices (“latex beads”, lipid vesicles, whole chromosomes, cells and biomolecules including proteins and DNA, as well as metal or semiconductor colloids and clusters.
Important for the applications to be described is the fact that the flow-mediated attractive interaction between beads extends to distances far exceeding the characteristic bead dimension. Planar aggregates are formed in response to an externally applied electric field and disassemble when the field is removed. The strength of the applied field determines the strength of the attractive interaction that underlies the array assembly process and thereby selects the specific arrangement adopted by the beads within the an-ay. That is, as a function of increasing applied voltage, beads first form planar aggregates in which particles are mobile and loosely packed, then assume a tighter packing, and finally exhibit a spatial arrangement in the form of a crystalline, or ordered, array resembling a raft of bubbles. The sequence of transitions between states of increasing internal order is reversible, including complete disassembly of planar aggregates when the applied voltage is removed. In another arrangement, at low initial concentration, beads form small clusters which in tun assume positions within an ordered “superstructure”.
II. Patterning of Silicon Oxide Electrode Surfaces
Electrode patterning in accordance with a predetermined design facilitates the quasi-permanent modification of the electrical impedance of the EIS (Electrolyte-Insulator-Semiconductor) structure of interest here. By spatially modulating the EIS impedance, electrode-patterning determines the ionic current in the vicinity of the electrode. Depending on the frequency of the applied electric field, beads either seek out, or avoid, regions of high ionic current. Spatial patterning therefore conveys explicit external control over the placement and shape of bead arrays.
While patterning may be achieved in many ways, two procedures offer particular advantages. First, UV-mediated re-growth of a thin oxide layer on a properly prepared silicon surface is a convenient methodology that avoids photolithographic resist patterning and etching. In the presence of oxygen, UV illumination mediates the conversion of exposed silicon into oxide. Specifically, the thickness of the oxide layer depends on the exposure time and may thus be spatially modulated by placing patterned masks into the UV illumination path. This modulation in thickness, with typical variations of approximately 10 Angstroms, translates into spatial modulations in the impedance of the Si/SiOx interface while leaving a flat and chemically homogeneous top surface exposed to the electrolyte solution. Second, spatial modulations in the distribution of the electrode surface charge may be produced by UV-mediated photochemical oxidation of a suitable chemical species that is fist deposited as a monolayer film on the SiOx surface. This method permits fine control over local features of the electrode double layer and thus over the electrokinetic flow.
A variation of this photochemical modulation is the creation of lateral gradients in the EIS impedance and hence in the current generated in response to the applied electric field. For example, this is readily accomplished by controlling the UV exposure so as to introduce a slow lateral variation in the oxide thickness or in the surface charge density. As discussed below, control over lateral gradients serves to induce lateral bead transport and facilitates the implementation of such fundamental operations as capturing and channeling of beads to a predetermined destination along conduits in the form of impedance features embedded in the Si/SiOx interface. Photochemical patterning of functionalized chemical overlayers also applies to other types of electrode surfaces including ITO.
III. Light-controlled Modulation of the Interfacial Impedance
The spatial and temporal modulation of the EIS-impedance in accordance with a pattern of external illumination provides the basis to control the electrokinetic forces that mediate bead aggregation. The light-modulated electrokinetic assembly of planar colloidal arrays facilitates remote interactive control over the formation, placement and rearrangement of bead arrays in response to corresponding illumination patterns and thereby offers a wide range of interactive manipulations of colloidal beads and biomolecules.
To understand the principle of this methodology, it will be helpful to briefly review pertinent photoelectric properties of semiconductors, or more specifically, those of the EIS structure formed by the Electrolyte solution (E), the Insulating SiOx layer (I) and the Semiconductor (S). The photoelectric characteristics of this structure are closely related to those of a standard Metal-Insulator-Semiconductor (MIS) or Metal-Oxide-Semiconductor (MOS) devices which are described in S. M. Sze, “The Physics of Semiconductors”, 2nd Edition, Chapt. 7 (Wiley Interscience 1981), the contents of which are incorporated herein by reference.
The interface between the semiconductor and the insulating oxide layer deserves special attention. Crucial to the understanding of the electrical response of the MOS structure to light is the concept of a space charge region of small but finite thickness that forms at the Si/SiOx interface in the presence of a bias potential. In the case of the EIS structure, an effective bias, in the form of a junction potential, is present under all but very special conditions. The space charge region forms in response to the distortion of the semiconductor's valence and conduction bands (“band bending”) in the vicinity of the interface. This condition in turn reflects the fact that, while there is a bias potential across the interface, there is ideally no charge transfer in the presence of the insulting oxide. That is, in electrochemical language, the EIS structure eliminates Faradaic effects. Instead, charges of opposite sign accumulate on either side of the insulating oxide layer and generate a finite polarization.
In the presence of a reverse bias, the valence and conduction band edges of an n-doped semiconductor bend upward near the Si/SiOx interface and electrons flow out of the interfacial region in response to the corresponding potential gradient. As a result, a majority carrier depletion layer is formed in the vicinity of the Si/SiOx interface. Light absorption in the semiconductor provides a mechanism to create electron-hole pairs within this region. Provided that they do not instantaneously recombine, electron-hole pairs are split by the locally acting electric field, and a corresponding photocurrent flows. It is this latter effect that affords control over the electrokinetic assembly of beads in the electrolyte solution.
To understand in more detail the pertinent frequency dependence of the light-induced modulation of the EIS impedance, two aspects of the equivalent circuit representing the EIS structure are noteworthy. First, there are close analogies between the detailed electrical characteristics of the electric double layer at the electrolyte-oxide interface, and the depletion layer at the interface between the semiconductor and the insulator. As with the double layer, the depletion layer exhibits electrical characteristics similar to those of a capacitor with a voltage-dependent capacitance. As discussed, illumination serves to lower the impedance of the depletion layer. Second, given its capacitive electrical response, the oxide layer will pass current only above a characteristic (“threshold”) frequency. Consequently, provided that the frequency of the applied voltage exceeds the threshold, illumination can lower the effective impedance of the entire EIS structure.
This effective reduction of the EIS impedance also depends on the light intensity which determines the rate of generation of electron-hole pairs. In the absence of significant recombination, the majority of photogenerated electrons flow out of the depletion region and contribute to the photocurrent. The remaining hole charge accumulates near the Si/SiOx interface and screens the electric field acting in the depletion region. As a result, the rate of recombination increases, and the efficiency of electron-hole separation, and hence the photocurrent, decreases. For given values of frequency and amplitude of the applied voltage, one therefore expects that as the illumination intensity increases, the current initially increases to a maximum level and then decreases. Similarly, the impedance initially decreases to a minimum value (at maximum current) and then decreases.
This intensity dependence may be used to advantage to induce the lateral displacement of beads between fully exposed and partially masked regions of the interface. As the illumination intensity is increased, the fully exposed regions will correspond to the regions of interface of lowest impedance, and hence of highest current, and beads will be drawn into these regions. As the fully exposed regions reach the state of decreasing photocurrent, the effective EIS impedance in those regions may exceed that of partially masked regions, with a resulting inversion of the lateral gradient in current. Beads will then be drawn out of the fully exposed regions. Additionally, time-varying changes in the illumination pattern may be used to effect bead motion.
IV. Integration of Biochemical Analysis in a Miniaturized, Planar Format
The implementation of assays in a planar array format, particularly in the context of biomolecular screening and medical diagnostics, has the advantage of a high degree of parallelity and automation so as to realize high throughput in complex, multi-step analytical protocols. Miniaturization will result in a decrease in pertinent mixing times reflecting the small spatial scale, as well as in a reduction of requisite sample and reagent volumes as well as power requirements. The integration of biochemical analytical techniques into a miniaturized system on the surface of a planar substrate (“chip”) would yield substantial improvements in the performance, and reduction in cost, of analytical and diagnostic procedures.
Within the context of DNA manipulation and analysis, initial steps have been taken in this direction (i.e., miniaturization) by combining on a glass substrate, the restriction enzyme treatment of DNA and the subsequent separation of enzyme digests by capillary electrophoresis, see, for example, Ramsey, PCT Publication No. WO 96/04547, the contents of which are incorporated herein by reference, or the amplification of DNA sequences by application of the polymerase chain reaction (PCR) with subsequent elecrephoretic separation, see, for example, U.S. Pat. Nos. 5,498,392 and 5,587,128 to Wilding et al., the contents of which are incorporated herein by reference.
While these standard laboratory processes have been demonstrated in a miniaturized format, they have not been used to form a complete system. A complete system will require additional manipulation such as front-end sample processing, binding and functional assays and the detection of small signals followed by information processing. The true challenge is that of complete functional integration because it is here that system architecture and design constraints on individual components will manifest themselves. For example, a fluidic process is required to concatenate analytical steps that require the spatial separation, and subsequent transport to new locations, of sets of analyte. Several possibilities have been considered including electroosmotic pumping and transport of droplets by temperature-induced gradients in local surface tension. While feasible in demonstration experiments, these techniques place rather severe requirements on the overall systems lay-out to handle the very considerable DC voltages required for efficient electroosmotic mixing or to restrict substrate heating when generating thermally generated surface tension gradients so as to avoid adverse effects on protein and other samples